Re: Heterodoxers are crackpots

["John M. Legge" <jlegge@xxxxxxxxxxxxxx>]

See interspersed responses

-----Original Message-----
From: Paul Davidson [mailto:pdavidson@xxxxxxx]
Sent: Wednesday, 4 September 2002 10:32 AM

 

<snip> 

There is no such thing as infinite precision!  All valid statistical predictions come with an associated standard error! 

 

 And that error will increase with time bounded by the appropriate Lyapunov exponent.  For a system to be ergodic the Lyapunov exponent would have to be zero or negative, with prediction and reality converging; but for all known complex systems the exponent is positive and the errors in predictions increase with time.

<snip> 

 even for next year, since practically all of them involve either limited
liability companies or sovereign states, either of whom can dishonour
contracts by citing force majeure.

They can cite what they want, but they must prove it in a court of law or else the State will enforce compliance -- or a payment in monetary damages.!  That much is certain!

 Had my retirement savings a year ago been
in Argentine bonds and Enron energy futures, insured by HIH Ltd, I would
probably not be sleeping very soundly tonight.

 

HIH was a limited liability insurance company with whom many Australians and Canadians had executed insurance contracts.  When HIH got into difficulty it appointed receivers who voided all outstanding insurance contracts perfectly legally.  Enron sold future contracts to deliver energy; again, faced with losses it invoked Chapter 11 and voided the contracts, perfectly legally.

 

Since 1846 in England, and only a little later in the US limited liability has become a standard feature of incorporation such that creditors have no recourse to shareholders in the event of bankruptcy.  In Australia ASIC, the SEC equivalent, does not even take action against directors of failed companies as long as creditors get at least 50 cents in the dollar and there are no smoking guns exposed to public view.  US practice is, I understand, similar.  Put brutally, half the value of any contract with a corporation can be voided at the will of its directors without any penalty.  The directors can void the other half if they have preserved enough of the company's cash in their private accounts to buy lawyers capable of persuading a jury that acts of outright fraud were merely acts carried out in good faith by persons of honest, if flawed, judgement.

 
So what does that tell you - namely that asset prices are NOT  generated by an ergodic stochastic process.  But we do have institutions built in to financial markets that suppose to maintain orderliness in the changing of asset prices.


 <snip> 

 OTH if Bruce's entrepreneur was considering buying a country
franchise for Coca Cola he/she could do so with very considerable confidence
(plus or minus one percent at the one sigma level) in next year's product
sales.

Is that sigma statistically determined?

 

The rate of change of market share for fast moving consumer goods such as Coca Cola has been intensively studied over a period of at least sixty years by people familiar with advanced statistical tools.  Year on year changes are generally less than one per cent,

 

JML

 

Paul Davidson